Give Claude Code real token-cost awareness.
This package ships a Claude Code skill
(tokenometer-cost-check) that teaches Claude Code agents to invoke the
tokenometer CLI for empirical prompt-cost measurement. With it
installed, Claude Code can answer "what does this prompt cost?" with a
real number sourced from provider countTokens APIs and the
tokenlens pricing registry,
instead of guessing from tiktoken.
One-shot copy:
cp -R packages/claude-code-skill ~/.claude/skills/tokenometerOr, for a one-file install:
mkdir -p ~/.claude/skills/tokenometer \
&& cp packages/claude-code-skill/SKILL.md ~/.claude/skills/tokenometer/Or use the included installer:
./packages/claude-code-skill/install.shThat writes ~/.claude/skills/tokenometer/SKILL.md. Restart Claude Code
(or open a new session) and the tokenometer-cost-check skill becomes
discoverable.
The skill registers trigger phrases like "what does this prompt cost",
"is this cheaper as JSON or YAML", "is GPT-4o or Claude faster on this
prompt", and "did my change increase prompt cost". When matched, Claude
Code knows to shell out to npx tokenometer … instead of guessing —
then surfaces real token counts, USD cost, and (with --latency) TTFT
- tokens/sec across Claude, GPT-4o, Gemini, Mistral, and Cohere, with an honesty flag for approximate vs exact counts.
See SKILL.md for the full skill content (trigger
phrases, invocation patterns, output handling, and the empirical
findings the skill is allowed to cite).
We'll submit this to the community skills registry once one stabilizes; track issue [#TBD]. Until then, install via the copy / installer commands above.
MIT